Frequently Asked Questions

Find answers to common questions about using the platform, accessing datasets, and other resources.

CORPORATE SOCIAL RESPONSIBILITY

Whether the ‘average net profit’ criteria mentioned under section 135(5) is Net profit before tax or Net profit after tax?

Computation of net profit for section 135 is as per section 198 of the Companies Act, 2013 which is primarily PROFIT BEFORE TAX (PBT) with other adjustments as referred in Rule 2(h) of CSR Rules. [Clarified by MCA vide General Circular No. 1/2016 dated 12th January, 2016]

Microsoft top questions

What is the difference between Microsoft 365 (subscription) and Office 2024 (one-time purchase)?

Microsoft 365 is a subscription that includes the most collaborative, up-to-date features in one seamless, integrated experience. Microsoft 365 includes the robust Office desktop apps that you’re familiar with, like Word, PowerPoint, and Excel. You also get extra online storage and cloud-connected features that let you collaborate on files in real time. With a subscription, you'll always have the latest features, fixes, and security updates along with ongoing tech support at no extra cost. You can choose to pay for your subscription on a monthly or yearly basis, and the Microsoft 365 Family plan lets you share your subscription with your family for up to 6 people, and use your apps on multiple PCs, Macs, tablets, and phones2.

Office 2024 is sold as a one-time purchase, which means you pay a single, up-front cost to get productivity apps for one computer. One-time purchases are available for both PCs and Macs. However, there are no upgrade options which means if you plan to upgrade to the next major release, you'll have to buy it at full price.

How do I know if my PC or Mac can run Microsoft 365?

Microsoft 365 is compatible with PC, Mac, Android, and iOS. See system requirements for compatible versions of your devices, and for other feature requirements.

test faq

Interactive nature:

Tutorials can involve questions, exercises, or discussions to engage the learner and reinforce learning.

test

Getting Started

What is AIKosha?

A platform that provides repository of datasets, models and use cases to enable AI innovation. It also features AI sandbox capabilities through an integrated development environment along with tools and tutorials.

What roles are available on the platform?

Roles include Explorers, Contributors, and Organization Administrators, each with specific privileges as below.

Explorer: Explorers have view-only rights on the platform. They can browse and download artefacts (dataset, model, or other resources) but can’t make any modifications or contribute new content.
Contributor: Contributors have all the rights of the explorer. In addition, they have permission to upload, modify, and manage artefacts on the platform. They can also approve or restrict access to the artefacts they contribute.
Organization Administrators (Admins): Admins have all the rights of an Explorer and a Contributor. In addition, they can approve/reject users request to join organization, manage artefacts published by the organization members and approve / reject publish and deletion requests for different artefacts. They can also promote / demote user roles within the organization.

What are the usage rights for the artefacts on the platform?

Each artefact (dataset, model, or other resource) in AIKosha is governed by specific permission settings that define usage rights. The available options are as below:

Open: Dataset will be Discoverable (searchable) and Downloadable by all the registered users on the platform
Restricted: Dataset will be discoverable (searchable) by all the users on the platform but require explicit approval from the contributor before they can be downloaded.
Private: Dataset will be downloadable by the contributor only.

testing FAQ niketa

In response to the data onboarding, guidance regarding the mechanism that AIKosh will adopt for extracting the data.

AIkosh enable user to upload the data using following routes:

Direct Upload: Platform enables upload of files by user from their system (recommended for small to medium sized datasets). In this mode, data loaded with AIKosh. The protocol used here is HTTP.
Push API: Contributors send datasets securely to AIKosh via REST API exposed by AIKosh. In this case data loaded with AIKosh. The protocol used here is HTTP.
Pull API: AIKosh retrieves dataset metadata from contributor REST APIs developed using AIKosh specifications. The data will remain with the contributor and will be downloaded directly to the explorers system through APIs. The protocol used here is HTTP.
SFTP: Platform enables upload of large files securely using temporary SFTP credentials. Users need to connect to AIKosh SFTP endpoint through their preferred SFTP client and using the credentials provided. In this case also the data loaded with AIKosh. The protocol used here is SFTP.
Public Platform: Allow users to contribute artefacts via public platforms — fetch datasets from Kaggle via SDK, pull from public GitHub repos using repo URLs, or link datasets directly from Hugging Face using URLs.

Kindly note: API and SFTP-based ingestion are recommended depending on data size and refresh requirements. In the future, we can also explore setting up dedicated data pipelines for streamlined integration. For more information, you can refer User Manual attached in the email.

The Quantum of data required for AIKosh.

The data available on AIKosh can be leveraged by the AI community to develop solutions aligned with specific use cases. While larger volumes of data generally contribute to better model performance, the exact amount needed can vary significantly depending on the nature of the use case. For effective model development, higher data volumes are typically preferable. However, starting with a representative variety of data could also be important. Capturing this variety in the dataset metadata ensures transparency and supports more robust AI training from the outset. Hence we encourage entities to share non-personal data and/or anonymized data.

Frequency of dataset needs to be uploaded.

The update frequency of a dataset depends on its nature and intended use. If no further updates are expected, contributors can mark the dataset as static using the frequency metadata field. For datasets

requiring periodic updates, contributors are expected to ensure timely revisions. In such cases, API-based ingestion is recommended, as it allows automatic updates on AIKosh whenever the source data is

refreshed.

As the data and domain owner, you are best positioned to determine the appropriate update frequency and provide guidance on data relevance and structure. At this stage, we are looking to onboard anonymized data sets across domains.

Data Standards including Meta data are to be prescribed.

In 2011, MeitY had set up an institutional mechanism for evolving / adopting standards for e-governance

applications to promote use of open standards. To this end, it developed Metadata and Data Standards

(https://docs.apisetu.gov.in/document-central/mdds/) such as One Catalogue (OneCat) and Data Exchange (DeCat), in line with global standards, to ensure interoperability, uniformity, and seamless exchange of data across systems. By setting guidelines for developing standards, MeitY aims to establish a framework for data management and security that helps standardize data management across various

government entities and promote transparency and accountability in data governance.

For a detailed description of the metadata fields used in the datasets, including the dataset name, licensing information, and other key attributes, please refer to the link: ------------------

The tool for anonymization of data.

The AIKosh toolkit section includes a selection of opensource tools that users can run on their local systems, including a tool for data anonymization. While these tools are made available to support experimentation and development, AIKosh does not endorse or certify their outputs. Please note that our anonymization tool currently in developing phase supports structured data only, and we strongly recommend that concerned division/users validate the output to ensure it aligns with their specific requirements, particularly when working with sensitive or regulated information.

Different domains may find certain tools more effective than others. Users are encouraged to evaluate the tools applicability/accuracy for their specific use cases and validate any results before making data-driven decisions. Your IT team or implementation partners can help identify which tools work best for your needs.

Presently the consent of data principle has not been taken for sharing the data then how the data can be shared.

At present, organizations processing (collecting, retaining, sharing) personal or sensitive personal data or information ought to comply with the Sensitive Personal Data or Information (SPDI) Rules of 2011 framed under the Information Technology Act of 2008, along with sectoral laws, if any. Please note that the SPDI Rules of 2011 are framed in relation to personal or sensitive personal data or information - and do not concern non-personal data or anonymous data leveraged by the AIKosh platform and the Open Government Data (OGD) platform. As you know, ministries have been engaging with the OGD platform for several years now.

While the OGD platform allows for free and open access to government-owned shareable data, the AI Kosh platform (as detailed above in item (i)(a)) restricts access to data leveraged through its platform:

1. Only verified / registered users can access the Artefact (datasets, models etc).

2. Contributor has the right to allow or disallow access to the Artefacts contributed by it under the restricted access mode.

Further, we ensure that all registered users (whether govt. or non-govt.) of the AIKosh platform comply with the terms of use and privacy policy before engaging with the platform.

How the data is to be processed for submission in respect to the DPDP Rule 2025 and legal perspective/checks related thereto.

Regarding applicability of the Digital Personal Data Protection Act of 2023 (DPDP Act), it is highlighted that the said act is not in force currently and pertains to processing of digital personal data. Non personal data is outside the ambit of DPDP Act.

It is reiterated that AIKosh is a platform offering services and products (which may include datasets, models, applications, etc) in the nature of non-personal data / anonymized form. Hence, you are advised to anonymize the data (such that it removes identifiable attributes from a data value to prevent identification of a person and renders the data truly anonymous) you intend to share on the AIKosh platform.

The AIKosh platform allows organizations (both government and non-government entities) to onboard

the platform as “contributors” permitting them to upload Artefacts (i.e. models, datasets, etc) on the

platform in the nature of anonymized data / nonpersonal data.

Contributors have the option to restrict access to the Artefacts they contribute to the AIKosh platform, allowing them to vet/ review the applicants before allowing them access to their Artefacts.

Format in which the data will be shared on AIKosh.

Structured Formats: CSV, TSV, Parquet, XLSX,

● Semi Structured Format: JSON, XML, YAML

● Unstructured Formats: PDF, TXT, MD, Sound (WAV, MP3, FLAC, AAC), Video (MP4, MKV, AVI), Image (JPEG, PNG, TIFF, BMP, RAW)

What is Data Anonymization.

Anonymization refers to the process of removing or modifying personal or sensitive data to ensure that individuals cannot be identified. As this is a data-specific process, it is best handled by the data contributor. Given that the Ministry is more familiar with the nature and sensitivity of their data, we recommend that contributors anonymize the data at their end before uploading it to AIKosh.

With reference to Personally Identifiable Information (PII), since data extraction is to be carried out from the real-world environment, which PII fields need to be excluded from the dataset contributed to AIKosh.

As you know, the AIKosh platform is curated for offering products and services basis non-personal data and/or anonymised data. Hence, prior to sharing data with us for the purposes of AIKosh, you are advised to remove identifiable attributes from a data value to prevent identification of a person and renders the data truly anonymous.

All applicable use cases under the Digital Personal Data Protection (DPDP) act should be taken into consideration while processing final PII fields removal from the requested data.

Registration and Login

How do I register on AIKosha?

Click on the "Register" button on the homepage, choose Individual registration, and click on sign-up (Government officials can also register via Parichay). Verify your phone number via an OTP and provide the necessary profile details to complete registration. This will register the users as explorer.

How do I register an organization on AIKosha?

Every user needs to first register as an explorer on the platform. The organization registration is then facilitated by Entity Locker. If the organization is already registered with Entity Locker, on the Explorer Dashboard, you can raise a request to create an organization by clicking on “Raise request”. Then log in via Entity Locker using Aadhar credentials and proceed to provide organization details to gain access to the Organization Admin Dashboard.

If the organization is not yet registered with Entity Locker, an authorized representative must first complete the Entity Locker registration. Upon approval, the entity is onboarded to Entity Locker. The user then returns to AIKosha and registers as an organization as explained in first paragraph.

For a step-by-step guide on the registration process, please refer the user manual on the platform.

What information do I need to provide during registration?

For individuals, you need to provide your name, email, and phone number. For organizations, you need to provide entity type, sector, organization name, website link, registered address, and company logo (optional).

How do I become a contributor on AIKosha?

On the Explorer Dashboard, click on “raise a request” to join an organization. You should enter necessary details like organization name, designation and employee ID. You will receive contributor rights subject to approval from organization admin. Ensure that your organization is registered on the platform beforehand.

Can I register without belonging to an organization?

Yes, you can register as an individual Explorer.

How do I get to know when my organization registration request on Entity Locker is approved?

You will be notified by DIC team via email once approved.

How can I log in on AIKosha?

Click on the "Login" button on the homepage. You can login via your phone number or username or government ID and an OTP verification.

What entity types are supported for registration?

Public Organizations, Private Companies, Non-profit organizations, Research Organizations, Universities and Startups/MSMEs are supported for registration.

Datasets

Can I contribute my own datasets to the platform?

Only contributors are allowed to upload datasets on AIKosha, subject to verification from the organization admin and platform admin.

How can I upload a dataset?

Ensure you are a contributor of an organization. Navigate to the Datasets page on the navigation pane and click on "Upload Dataset". Upload the dataset, fill in metadata and set access permissions. The dataset will be published on the platform subject to verification from the organization admin and platform admin.

What file formats are supported for datasets?

Supported formats include CSV, TSV, Parquet, Excel, JSON, XML, YAML, PDF, Sound Files (WAV, MP3, FLAC, AAC), Video Files (MP4, MKV, AVI), Image Format (JPEG, PNG, TIFF, BMP, RAW), Text (TXT), or Markdown file format (MD).

Can I update or modify my uploaded dataset?

Yes, contributors can modify metadata and update new versions of the data files.

How do I restrict access to my dataset?

Contributors can set access controls to define how users can discover and download their dataset. Options include making it open, restricted and private.

Are there any usage restrictions on the datasets?

Explorers can view and download open datasets directly without raising any requests. For restricted datasets, users must submit a request stating the reason for download. The dataset owner will review the request and decide on granting access.

Models

Can I contribute my own models on AIKosha?

Only contributors are allowed to upload models on AIKosha, subject to verification from the organization admin and platform admin.

How can I upload a model?

Ensure you are a contributor of an organization. Navigate to the Models page and click on "Upload Model". Upload the model, fill in metadata and set access permissions. The model will be published on the platform subject to verification from the organization admin and platform admin.

Can I update or modify my uploaded model?

Contributors can modify metadata for the model.

How do I restrict access to my model?

Contributors can set access controls to define how users can discover and download their model. Options include making it open, restricted and private.

Can I change the access settings for my uploaded model?

Yes, you can modify the access control of your model to open, restricted, or private at any time.

What types of models can I upload?

Supported formats include Python Script (PY), Compressed Archive (ZIP, Z, GZ, BZ2, XZ, LZMA), JSON and Machine Learning Model (SafeTensors, PT, ONNX, HDF5, PKL), TSV.

Test FAQ

On which companies, Corporate Social Responsibility Policy (CSR) provisions specified under section 135 of the Companies Act, 2013 are applicable?

A company satisfying any of the following criteria during the immediately preceding financial year has to comply with CSR provisions specified under section 135 of the Companies Act, 2013 read with the CSR Rules made thereunder: i. net worth of rupees five hundred crore or more, or ii. turnover of rupees one thousand crore or more or iii. a net profit (as calculated under Section 198 with other adjustments as referred in Rule 2(h) of CSR Rules.) of rupees five crore or more The Central Government may give directions to a company or class of companies as it considers necessary to ensure compliance of provisions of this section and such company or class of companies shall be required to comply with such directions

When does the Companies (Corporate Social Responsibility Policy) Amendment Rules, 2021 come into effect?

The Ministry of Corporate Affairs vide Notification No. G.S.R. 40(E) dated 22nd January, 2021 issued the Companies (Corporate Social Responsibility Policy) Amendment Rules, 2021. The same has been made effective from the date of their publication in the Official Gazette i.e. 22nd January, 2021.

Use Cases

What is a Use Case?

The Use Case Library is a repository showcasing real-world AI applications built using datasets and models from the platform.

How can I contribute a use case?

Contributors can submit use cases by providing detailed use-case descriptions, metadata, and associated datasets and models. For step-by-step upload process, you can refer the user manual on the platform.

Can I download use case resources?

Use cases include downloadable resources like datasets and models depending on the permissions set by the contributor of the respective artefacts.

User Management

How do organization admins manage user roles within an organization?

Admins can approve or reject requests from Explorers to join the organization. Administrators can manage user roles by assigning, upgrading, or downgrading them, and promote other organization members to Administrator roles. They also have the authority to delete or remove users and reassign ownership of artefacts from deleted users to other member of the organization.

What happens if my request to join an organization is denied?

The user will remain an Explorer on the platform but can re-apply for the same organization or join another organization.

Is it possible to transfer ownership of datasets or models?

Yes, ownership can be transferred to another user within the same organization by the organization admin.

How can I track my requests to join an organization as a contributor or download restricted artefacts?

The user can track all his requests in the “Recent Activities” section in the navigation pane. You will also be notified of the approval / rejection on email (along with reason in case of rejection).

Can Organization Admin edit or delete datasets / models uploaded by contributors?

Yes, as an Organization Admin, you have the authority to edit metadata, change visibility, or delete datasets and models uploaded by contributors in your organization.

Can contributor upload datasets/models without admin approval?

All uploaded datasets and models must be approved by the Organization Admin and Platform Admin before they are visible on the platform.

Can I delete my account?

Yes, users can delete their accounts by going into the “Account settings” in My profile section.

What happens to contributors’ artefacts uploads if they delete their account?

The contributor must assign the ownership of artefacts to another contributor in his organization or his organization admin before deleting the account.

Grievance Support

How do I raise a grievance?

Navigate to the “Help Centre” in your profile and go to "Report Grievances" page. Fill out the form with the required details and submit it. You will receive a tracking number to track your grievance. You can also write to us on email for assistance or contact us on the support number provided.

How do I track my grievance?

Users will receive a tracking number on the platform as well as on email to track their grievance. You can track the status on “Report Grievances” section under Help Centre.

Security Mechanism

What security measures are implemented on AIKosha?

Security is a top priority at AIKosha.The platform utilizes National Single Sign-On (SSO) for login and authentication.Role-based and permission-based access controls are implemented to regulate access to artefacts. Additionally, encryption is applied to both data at rest and data in transit. Advanced security settings are enabled in the firewall to prevent malicious activities, protecting the portal and its content.

Accessibility options by UX4G

AIKosh

Resources

Support